PROJECT SUMMARY Alzheimer's disease (AD) is the most common late life dementia and affects approximately 6 million Americans, therefore creating a huge social and economic impact. More importantly, the persistent demographic shift to- wards an older population will cause the number of AD patients to double within the next 20 years. Clinically, AD is de?ned by progressively worsening memory loss, cognitive decline, behavioral changes, and ultimately death. Pathophysiologically, AD is characterized by the gradual accumulation of toxic protein deposits that spread through the brain and eventually result in wide-spread neuron cell death and cerebral atrophy (CA). However, de- spite signi?cant advances in our understanding of pathophysiology in AD and related dementias (ADRD), neither a de?nitive, antemortem diagnostic tool nor a pharmacological cure exist today. The early detection of AD and ADRD has proven particularly challenging because the biological processes most often precede the onset of clinical symptoms by up to two decades and, therefore, progress unnoticed during a time at which intervention is considered to be most effective. Thus far, ?ve biomarkers have been developed to visualize established AD hallmark features: toxic deposits of ?-amyloid and tau proteins and neurodegeneration associated with cortical thinning and brain volume loss. These biomarkers are invasive and resource intensive measures, however, and involve the exposure to radioactive tracers in amyloid and tau PET or a lumbar puncture for CSF immunoassays. In this project we propose a novel mechanobiological disease model for AD which predicts the prion-like protein progression and subsequent structural changes on the organ-level in space and time. Our com- putational approach utilizes medical images and physics-based modeling to provide subject-speci?c simulations of these common features in AD with the goal to minimize exposure to invasive measures. We hypothesize that our model is a reliable biomarker to enable earlier diagnosis of dementia type and monitoring of disease pro- gression which would allow for the development of more effective and personalized treatment strategies. To test our hypothesis, we will use longitudinal biomarker data (amyloid/tau PET and MRI) from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and validate our model in ten subjects. For each subject, we will reconstruct their brain in a computer model, use their initial PET scans to calibrate our AD progression model, and compare our subsequent numerical predictions of biomarker progression against their follow up scans. This approach aims at integrating existing technologies that visualize temporal and spatial patterns of individual biomarkers into a noninvasive disease model for AD and ADRD. By capturing the fundamental mechanisms of AD and ADRD, we can, for the ?rst time, systematically study organ-level features of individual dementias. As such, this study is particularly relevant to public health because early diagnosis of dementia type and a reliable tool to track disease progression will have a big clinical impact on disease management and minimize frequent expo- sure to invasive biomarkers. Demonstrating the role of mechanobiology in AD pathology will inherently advance our basic understanding of other neurodegenerative diseases, such as Parkinson's disease, amyotrophic lateral sclerosis (ALS), and chronic traumatic encephalopathy (CTE).